System Requirements#
This page lists the hardware, software, and platform requirements for running the Agentic Predictive Maintenance (APM) blueprint.
Hardware Platforms Used for Validation#
4th and 5th Gen Intel® Xeon® processors
Intel® Arc™ GPU (A-series and B-series) with compatible Intel® Xeon® processor or Intel® Core™ processor
Intel® Core™ Ultra processors with integrated GPU (suitable for smaller pipelines and fallback mode)
Large Language Model (LLM) mode requires GPU support. The fallback mode (rule-based) runs on CPU alone.
Operating Systems Used for Validation#
Ubuntu OS version 22.04 LTS for CPU-only configurations
Ubuntu OS version 24.04 LTS when using discrete GPU hardware
See the Intel GPU driver documentation for specific kernel requirements per GPU model.
Minimum Hardware Configuration#
Component |
Minimum |
Recommended |
|---|---|---|
RAM |
16 GB |
32 GB |
Storage |
50 GB |
100 GB |
CPU |
Any Intel® Xeon® processor or Intel® Core™ processor |
4th Gen Intel® Xeon® processor or later |
GPU |
Not required (fallback mode) |
Intel® Arc™ GPU for LLM mode |
Software Requirements#
Software |
Version |
|---|---|
Docker Engine |
24.0 or higher |
Docker Compose tool |
2.20 or higher |
Python programming language |
3.10 or higher (for data preparation only) |
Compatibility Notes#
Known Limitations:
The LLM service offers Neural Processing Unit (NPU) inference as an experimental option, but the option is not validated for all model and configuration combinations.
Intel® Core™ Ultra processors Series 2 and 3 with integrated GPU can run the application but model selection significantly affects performance.
Validation#
Ensure Docker Engine and Docker Compose tool are installed and are running before following the Get Started section.